Unverified50% confidenceFactExact time
GPTQ is a layer-wise quantization method based on approximate second-order information, proposed by Frantar et al. in 2022, which minimizes quantization reconstruction error through the approximate inverse of the Hessian matrix.
1
Sources
50%
Confidence
Long-term
Relevance
9/6/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedGPTQ通过逆Hessian矩阵的Cholesky分解按列顺序量化权重并将量化误差分配到未量化权重上，该方法源自Optimal Brain Surgeon剪枝理论69% similarUnverifiedResearchers have found statistical anomalies in mainstream models such as GPT-4 and Claude, suggesting memorization rather than reasoning.61% similarUnverified支持向量机（SVM）由 Vapnik 等人在 1990 年代提出，通过核函数将低维线性不可分问题映射到高维空间求解59% similarUnverifiedmeshoptimizer 的网格简化算法基于二次误差度量（QEM）的变体实现，QEM 由 Garland 和 Heckbert 于1997年在 SIGGRAPH 上提出59% similarVerifiedAt the end of 2023, Terence Tao and collaborators used Lean and LLM assistance to prove an important special case of the polynomial Freiman-Ruzsa conjecture59% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/864125API
curl https://kongchang.com/api/v1/knowledge/claims/864125MCP
get_claim(id=864125)